Point to multipoint (P2MP) topology is one of the most commonly used topologies in an access network. In general, P2MP may include a root station (RS) and a number of leaf stations (LSs). In P2MP, any media having a RS that broadcasts packets through a single trunk (such as a frequency, wavelength, or wireless channel) to LSs typically may be referred to as downstream. Similarly, LSs unicasting packets through branches and the trunk to the RS may be referred to as upstream. In addition, the LSs may not communicate with each other in a peer-to-peer manner.
Many wired broadband access networks such as the Time Division Multiplex (TDM) Passive Optical Network (PON) (which includes Ethernet passive optical networks (EPONs), Gigabit passive optical networks (GPONs), and Broadband passive optical networks (BPONs)) can be generalized into a P2MP architecture. The P2MP architecture of PONs may reduce the dominant deployment and maintenance cost, and facilitates the central management by utilizing the RS as the central office.
In the recent past, there have been attempts to address upstream resource management and allocation mechanism issues in P2MP networks, especially in P2MP EPON networks. These schemes may be categorized into three categories: fixed resource allocation (FRA), request-based resource allocation (RRA), and prediction-based resource allocation (PRA). Although most of the schemes address the resource management in EPONs, they can be generalized to other P2MP networks by employing appropriate MAC control cells and fields in the frames.
Although attempts have been made to address the upstream resource management issue in P2MP networks, few attempts have addressed the above different resource management schemes such that these schemes can be evaluated, compared and further improved.
Furthermore, current upstream resource allocation schemes in P2MP networks may have difficulties reaching transient performance objectives such as minimum settling time and maximum overshoot, due to the complexity of mapping the objectives into the corresponding scheduling algorithm and resource management schemes.